Intratumoral injection of inactivated Sendai virus particles elicits strong antitumor activity by enhancing local CXCL10 expression and systemic NK cell activation

Abstract

We have already demonstrated that inactivated, replication-defective Sendai virus particles (HVJ-E) have a powerful antitumor effect by both the generation of tumor-specific cytotoxic T cells and inhibition of regulatory T cell activity. Here, we report that HVJ-E also has an antitumor effect through non-T cell immunity. Microarray analysis revealed that direct injection of HVJ-E induced the expression of CXCL10 in established Renca tumors. CXCL10 was secreted by dendritic cells in the tumors after HVJ-E injection. Quantitative real-time RT-PCR and immunohistochemistry revealed that CXCR3+ cells (predominantly NK cells) infiltrated the HVJ-E-injected tumors. Moreover, HVJ-E injection caused systemic activation of NK cells and enhanced their cytotoxity against tumor cells. In an in vivo experiment, approximately 50% of tumors were eradicated by HVJ-E injection, and this activity of HVJ-E against Renca tumors was largely abolished by NK cell depletion using anti-asialo GM1 antibody. Since HVJ-E injection induced systemic antitumor immunity by enhancing or correcting the chemokine-chemokine receptor axis, it might be a potential new therapy for cancer.

Fig. 1

Induction of CXCL10 by HVJ-E in vivo and in vitro. CXCL10 mRNA expression in whole tumors (a) or DCs isolated from tumors (b) after injection of HVJ-E or saline was measured by quantitative real-time RT-PCR (n = 5 per group). CXCL10 mRNA expression was increased by injection of HVJ-E. Error bars indicate the SE (<5%). This experiment was repeated three times with similar results. c CXCL10 levels in the medium of cultures with (+) or without (-) Renca cells, HVJ-E, or DCs. When HVJ-E and syngeneic mouse DCs were added to cultured Renca cells, a significant increase of CXCL10 production was observed. A significant increase of CXCL10 was also detected in the medium of DCs cultured with HVJ-E even in the absence of Renca cells, although the level was lower than that obtained with Renca cells (*P < 0.05, **P < 0.01). Data points are the mean ± SE of triplicate wells. SE < 5%. This experiment was repeated four times with similar results

Fig. 2

Infiltration and activation of NK cells after intratumoral injection of HVJ-E. a Intratumoral infiltration of immune cells was investigated by quantitative real-time RT-PCR (n = 5). DX5 mRNA expression showed a marked increase. SE < 5%. This experiment was repeated three times with similar results. b NK cell infiltration. Immunofluorescence staining using a FITC-conjugated anti-mouse CD49b (DX5)/Pan NK cell monoclonal antibody shows prominent infiltration of DX5-positive cells (green) into an HVJ-E-treated tumor (×200). This experiment was repeated three times with similar results. c Activation of intratumoral NK cells: NK cells were purified from tumors injected with HVJ-E or saline, and IFN-γ and CD69 mRNA expression was assayed by quantitative real-time RT-PCR. Both IFN-γ and CD69 expressions were upregulated in HVJ-E-treated tumors. SE < 5%. This experiment was repeated three times with similar results. d NK cytotoxity in vivo. Cytotoxicity assays were performed with NK cells harvested from the spleens of mice after treatment with HVJ-E (filled square) or saline (filled diamond). NK cells from HVJ-E-treated mice showed an increase of cytotoxity against Renca cells. This experiment was repeated three times with similar results

Fig. 3

Activation of NK cells by type I IFNs induced from HVJ-E-stimulated DCs. a Intratumor expression of type I IFN mRNAs after in vivo injection of HVJ-E or saline injection measured by quantitative real-time RT-PCR (n = 5/group). This experiment was repeated four times with similar results. b Type I IFNs in the conditioned medium of Renca cells cultured with HVJ-E in the presence or absence of DCs. HVJ-E was added at an MOI of 0.3–3,000. Both IFN-α and IFN-β levels were increased in an HVJ-E-dose-dependent manner (*P < 0.05, **P < 0.01) only in the cultures with DCs. Data points are the mean ± SE of triplicate wells. This experiment was repeated three times with similar results. c A significant increase of IFN-γ was observed in the culture medium of Renca cells after addition of H-DCCM and NK cells (*P < 0.01), while no significant increase of IFN-γ secretion was detected with either HVJ-E or H-DCCM alone. Data points are the mean ± SE of triplicate wells. This experiment was repeated three times with similar results. d IFN-γ secretion was reduced by anti-IFNAR2 in the medium of Renca cells cultured with H-DCCM. When NK cells were preincubated with anti-IFNAR2, the increase of IFN-γ in response to HVJ-E was abolished. Data points are the mean ± SE of triplicate wells. This experiment was repeated three times with similar results. SE < 5%. Results were statistically analyzed using the unpaired t test

Fig. 4

Suppression of the growth of Renca tumors in mice by intratumoral injection of HVJ-E. a Renca cells were inoculated intradermally into the backs of syngeneic BALB/c mice. Then HVJ-E (open circle) or saline (filled square) (n = 5 per group) was injected three times (on days 5, 10, and 15) into the resulting tumors. Tumor growth was significantly inhibited by HVJ-E injection (*P < 0.01) and approximately 50% of the mice became tumor-free. Data shown are representative of experiments that were repeated five times with similar results. b Kaplan–Meier survival curves for HVJ-E-treated and saline-treated mice. When HVJ-E (open circle) or saline (filled square) (n = 5 per group) was injected three times into the intradermal Renca tumors of BALB/c mice, the survival of HVJ-E-treated mice was significantly better than that of saline-treated mice (*P < 0.01). Data shown are representative of experiments that were repeated four times with similar results. c Loss of the antitumor effect of HVJ-E after neutralization of NK activity. Renca cells were inoculated intradermally into syngeneic BALB/c mice, and then HVJ-E was injected three times into the resulting tumors together with anti-asialo GM1 antibody (filled square) or control IgG (filled circle) (n = 5 per group). Tumor growth was inhibited in the mice treated with HVJ-E plus control IgG, whereas it was not inhibited in mice treated with HVJ-E plus anti-asialo GM1 antibody (*P < 0.05). Data shown are representative of experiments that were repeated three times with similar results. Arrows indicate the timing of injection. SE (<5%). Statistical analysis was done with the unpaired t test (a, c) and the Log-rank test (b)

Fig. 5

Induction of T cell immunity in Renca tumors in later phase. a Intratumoral infiltration of T cells 48 h after HVJ-E treatment was investigated by quantitative real-time RT-PCR (n = 3). CD8 and CD4 mRNA expressions showed a marked increase. These experiments were repeated three times with similar results. b ELISPOT analysis of HVJ-E-treated mice (open bars) and saline-treated mice (solid bars). Spleen cells were harvested from the mice at 7 days after the last injection of HVJ-E or saline into Renca tumors. Then the spleen cells were stimulated with MMC-treated Renca cells for 5 days, after which CD8+ T cells were isolated. Subsequently, 1 × 10⁵ purified CD8+ T cells were cultured for 48 h with or without MMC-treated Renca cells or CT26 cells and the number of IFN-γ-secreting CD8+ T cells was counted (*P < 0.05). Data points are the mean of triplicate wells. c Antitumor effect of HVJ-E in SCID mice (n = 5). Renca cells were inoculated intradermally into SCID mice, after which HVJ-E (open circle) or saline (filled square) (n = 5/group) was injected three times into the tumors that developed. Tumor growth was significantly inhibited by HVJ-E (*P < 0.05); although the inhibition was less marked compared with that in wild-type mice. Arrows indicate the timing of injection. This experiment was repeated twice with similar results. SE < 5%. Results were analyzed statistically using the unpaired t test (b, c)